Clarification regarding dmypy and "one set of source files"

I'm working to set up dmypy in a multi-root workspace where we'll need to run multiple instances of dmypy for various subsets of files. We're hoping to do something like dmypy --status-file dmypy_$subset.json run -- <flags> <subset files...> for each subset so they all get their own instance.

However, I'm curious how the server will deal with the addition of new files to one of the subsets in light of this note on the readme:

Each mypy daemon process supports one user and one set of source files, and it can only process one type checking request at a time. You can run multiple mypy daemon processes to type check multiple repositories.

For example, if we typically do

dmypy --status-file .dmypy_foo.json run -- <flags> foo.py bar.py

what happens when we add some baz.py to the repo? We'd start calling

dmypy --status-file .dmypy_foo.json run -- <flags> foo.py bar.py baz.py

Is that going to break anything? Do we need to restart the daemon and lose the caching benefits if there's a new file?

Start with the README passage describing one user and one set of source files, then inspect the dmypy command behavior for repeated runs with the same status file. Confirm what happens when baz.py is added to foo.py and bar.py, including whether the daemon must restart or can retain its cache, and document the result.